Bayesian Learning in Social Networks

Abstract

We study the (perfect Bayesian) equilibrium of a model of learning over a general social network. Each individual receives a signal about the underlying state of the world, observes the past actions of a stochastically-generated neighborhood of individuals, and chooses one of two possible actions. The stochastic process generating the neighborhoods defines the network topology (social network). The special case where each individual observes all past actions has been widely studied in the literature. We characterize pure-strategy equilibria for arbitrary stochastic and deterministic social networks and characterize the conditions under which there will be asymptotic learning-that is, the conditions under which, as the social network becomes large, individuals converge (in probability) to taking the right action. We show that when private beliefs are unbounded (meaning that the implied likelihood ratios are unbounded), there will be asymptotic learning as long as there is some minimal amount of "expansion in observations". Our main theorem shows that when the probability that each individual observes some other individual from the recent past converges to one as the social network becomes large, unbounded private beliefs are sufficient to ensure asymptotic learning. This theorem therefore establishes that, with unbounded private beliefs, there will be asymptotic learning in almost all reasonable social networks. We also show that for most network topologies, when private beliefs are bounded, there will not be asymptotic learning. In addition, in contrast to the special case where all past actions are observed, asymptotic learning is possible even with bounded beliefs in certain stochastic network topologies.Keywords: information aggregation, learning, social networks, herding, information cascades.JEL Classification: C72, D83. * We thank Lones Smith and Peter Sorensen for useful comments and suggestions. We gratefully acknowledge financial support from the AFOSR and the NSF.

Extensive Imitation is Irrational and Harmful*

“…We limit our analysis to rich-information settings in order to crisply articulate the observational conditions under which our form Yet neither that paper, nor related ones by Banerjee and Fudenberg (2004) and Smith and Sørensen (2008) that share the feature of Acemoglu et al (2011) that players only observe subsets of their predecessors, focus on how players' learning manifests itself in behavior. Not focusing on behavior short of the limit, these authors do not identify conditions determining whether rationality leads to anti-imitation.…”

Opinion Dynamics and Learning in Social Networks

“…Hence we have obtained the result that it is a strict best response for her to herd on the behavior of the first two agents, entirely ignoring her own information. 5 What about agent n = 4? Again, conditioning correctly on others' strategies, he will infer from x 1 = x 2 = 1 that s 1 = s 2 = 1, and also understand that agent 3 has ignored her information.…”

“…We will next present the model and the analysis from [5], which introduces a network structure to capture these interactions (as well as allowing more general signal structures). Allowing for a network structure leads to several challenges.…”

“…We now outline the model and the results from [5]. We continue to assume that the unknown state θ is binary, {0, 1} (and that P(θ = 0) = P(θ = 1) = 1/2).…”

Imitating the Most Successful Neighbor in Social Networks

“…Large part of this literaure has focused on Bayesian learning (see Gale and Kariv, 2003;Acemoglu et al, 2011) and best-response strategies.…”